Data-machine independence (DMI)
M. Botlo, M. Jagielski, Aguilar Romero, L. L. Miller · OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) · 1993
A distributed data acquisition system (DAQ) should not restrict the number or type of processors used in a given application. For any software applications that must operate on a multi-vendor distributed system, three serious problems have to be addressed regarding data handling: data representation, memory organization, and address alignment. Data representation, memory organization, and address alignment problems are inherent in the various hardware platforms available on the market today. Furthermore, issues such as how to coordinate parallel software efforts on multiple platforms that share data must be resolved. Traditionally such efforts are coordinated through meetings, memos, and documents describing shared data blocks (Data Interface Documents). The traditional approach is risky at best: the frequency of changes demanded by end-users is faster than the rate at which interface documents can be updated. Simple changes in data structures tend to dramatically increase the software development time, as all the tools must be modified to accommodate the changes. In Section 2.0 the authors examine in detail the technical implications for data handling when using distributed applications. In Section 3.0 they compare different approaches in handling the data in distributed environments. The four methods studied are: (1) Abstract Syntax Notation 1 (ASN.1), (2) Externalmore » Data Representation (XDR), (3) Self Describing Data Format (SDS), and (4) ZEBRA. They compared the four tools with the standard programming method (using {open_quotes}C{close_quotes}). In Section 4.0 they present conclusions of the test using the four different tools.« less